Unlock Editor’s Digest for free
FT editor Roula Khalaf has chosen her favorite stories in this weekly newsletter.
With U.S. beef prices hovering near record highs, North America’s largest beef processor is using AI to remove millions of dollars worth of excess meat from bones at its slaughterhouses.
Cargill has developed an AI-powered computer vision system that detects “red pixels,” tiny bits of meat stuck to bones and fat as a cow carcass moves through a processing line.
Early trials show that by using the CarVe brand system, meat packers can recover up to 0.5% more meat from each animal on average.
“This is huge considering the scale of our business,” said Florian Schattenmann, the company’s head of research and development.
Shattenman said AI is especially useful because beef processing involves variability. “It’s not like building a car. All Toyota Camrys look the same, but they’re all automated. Every cow is a little different.”
Minnesota-based Cargill said it produces about 11 billion pounds of beef annually. A 0.5 percentage point increase equates to approximately 55 million pounds of additional meat per year, or approximately $200 million at current wholesale beef prices.

The U.S. market produces more than 27 billion pounds of beef each year, and similar improvements across the industry could keep about 135 million pounds of extra beef in the food system, Cargill said.
CarVe works by scanning each carcass as it moves through the production line, identifying red meat and bones in real time.
Performance data is instantly displayed on a screen next to the cutting table, and each meat packer’s workstation is color coded red, yellow and green to indicate if there is too much meat on the carcass.
Supervisors will be able to play the clip on screen and guide their colleagues as best they can on the missing parts.
The system is currently being used at Cargill’s beef plants in Friona, Texas, and Fort Morgan, Colorado. The company said it also plans to introduce it at other meat packaging plants.
Cargill is also using AI to predict how many workers will show up for each shift at its meat packing plants, to increase efficiency in one of the most labor-intensive parts of the business.
The model leverages data from employee tenure to weather, holidays, and even local high school football schedules.
“It feeds an insane amount of internal and external data,” says Shattenman.
As the U.S. cattle herd shrinks to its smallest size in 75 years, some are looking for incremental gains. The nation’s herd numbered 86.2 million animals as of Jan. 1, according to the USDA’s latest inventory report, down from a year ago as ranchers decimated their herds after years of drought.
The contraction has caused beef prices to skyrocket, pushing retail beef prices to record levels, prompting steakhouses to reduce portion sizes and raise prices.
Ground meat prices rose about 17% in January from a year earlier, far outpacing general food inflation, according to government data. But protein-hungry consumers continue to buy beef even at higher prices, keeping demand resilient.

Cargill said it is investing heavily in technology upgrades to reduce waste and increase production per animal.
“By producing more cows from the same cow, CarVe improves resource efficiency, enhances supply, and helps manage cost pressures,” Cargill said, adding that the system “empowers” staff and is not designed to “oversee or replace people.”
